Switch to gpt-5-mini as the default

[ankrgyl]

No description provided.

[github-actions]

Braintrust eval report

‼️ Autoevals (gpt-5-1754668222)

Score	Average	Improvements	Regressions
NumericDiff	78.3% (+4pp)	20 🟢	9 🔴
Start	1754668221.91s	-	-
End	1754668229.86s	-	-
Duration	7.95s (+2.86s)	48 🟢	171 🔴
Llm_duration	9.92s (+5.62s)	-	119 🔴
Prompt_tokens	317.7tok (-209.2tok)	119 🟢	-
Completion_tokens	249.19tok (+203.62tok)	-	119 🔴
Total_tokens	566.89tok (-5.58tok)	72 🟢	47 🔴
Estimated_cost	0$ (0$)	119 🟢	-
Prompt_cached_tokens	0tok (+0tok)	-	-
Prompt_cache_creation_tokens	0tok (+0tok)	-	-

Expand to see errors

Error: Unknown score choice undefined
    at parseResponse (/home/runner/work/autoevals/autoevals/jsdist/index.js:430:11)
    at OpenAIClassifier (/home/runner/work/autoevals/autoevals/jsdist/index.js:406:10)
    at process.processTicksAndRejections (node:internal/process/task_queues:105:5)
    at async Object.ClosedQA (/home/runner/work/autoevals/autoevals/jsdist/index.js:456:12)
    at async /home/runner/work/autoevals/autoevals/node_modules/.pnpm/braintrust@0.0.140/node_modules/braintrust/dist/index.js:20435:22
    at async /home/runner/work/autoevals/autoevals/node_modules/.pnpm/braintrust@0.0.140/node_modules/braintrust/dist/index.js:20394:26
    at async Object.task (eval at <anonymous> (/home/runner/work/autoevals/autoevals/node_modules/.pnpm/braintrust@0.0.140/node_modules/braintrust/dist/cli.js:29191:5), <anonymous>:4191:20)
    at async rootSpan.traced.name (/home/runner/work/autoevals/autoevals/node_modules/.pnpm/braintrust@0.0.140/node_modules/braintrust/dist/cli.js:24576:26)
    at async callback (/home/runner/work/autoevals/autoevals/node_modules/.pnpm/braintrust@0.0.140/node_modules/braintrust/dist/cli.js:24572:11)
    at async /home/runner/work/autoevals/autoevals/node_modules/.pnpm/braintrust@0.0.140/node_modules/braintrust/dist/cli.js:24707:16